Money Stuff

Airline Shares and Whistle-Blowers

Also AI inscrutability, PCAOB scandal and bond market liquidity.


How do you feel about yesterday, if you're a United Continental Holdings shareholder? Well, the stock was down a lot, but then mostly got better, closing down just 1.13 percent on the day. United shareholders ended up losing about $257 million of equity market capitalization. But if you're actually a United shareholder, there's a pretty good chance you're also an American Airlines Group Inc. shareholder. According to Bloomberg, four of United's top five shareholders are also top-five holders of American. Three of them are also top-five holders of Delta Air Lines Inc. United's top 10 holders own about 49.8 percent of United Stock between them -- and about 51.6 percent of American and 37.6 percent of Delta. 

American was up 3.8 percent yesterday. Delta was also up, as was Southwest Airlines Co.; Alaska Air Group Inc. was down. (American was the biggest winner, perhaps because it is best positioned to gain market share from United -- or more likely because "a closely watched revenue measure rose more than expected in the first quarter," which "signaled progress for American in a two-year struggle by U.S. airlines to regain pricing power in the wake of a 2015 fare war.") If you owned all those airlines -- as the S&P 500 Index does -- in proportion to their market capitalizations, your airline holdings were up about 1 percent yesterday. Fortune did the math for Warren Buffett and found that he was up $104 million on airlines on the day, "illustrating once again the investor's talent for making money even when the headlines are grim." But of course if you owned the airlines through an S&P 500 fund, you demonstrated the same talent.

You know what I am getting at here. Bad news for United was easy for United's shareholders to shrug off, because United's shareholders are also shareholders in its competitors. To the extent the United fiasco leads to new airline regulations that crimp everyone's profit margins, that's bad for shareholders. To the extent that it leads to more people getting fed up with United and flying American instead, that doesn't matter to shareholders. "Boycotting United to fly another airline is just what United's index-fund owners want you to do," I joked on Twitter.

Is there causality here? It's a question we talk about a lot. It seems implausible that a gate agent in Chicago decided to summon police to beat up a passenger on the theory "well, this will be bad for us, but good for American Airlines, and our shareholders are mostly diversified quasi-indexers, so whatever." And yet there is a logic to it. Companies are supposed, in traditional corporate theory, to work on behalf of their shareholders. A lot depends on that supposition: Fining companies for misbehavior, for instance, only makes sense if you think that employees and managers want to avoid losing money for shareholders, since the employees are the ones who misbehave while the shareholders are the ones who pay the fines. If you don't think that the principal emergent desire of United, as a company, is to make its shareholders richer, then very few of United's actions will make sense. 

Meanwhile, United knows who its shareholders are. It can see that they mostly own other airlines too. If all it wants, deep in its corporate heart, is to make those shareholders richer, then why should it care whether they get richer on their United stock or their American stock? Why should it work hard to be the most-loved airline, if that's expensive and customer love is zero-sum? Why not focus on pricing power and cost-cutting, which makes shareholders richer, even if it might lead to a public-relations disaster for any particular airline? 

Elsewhere, Tyler Cowen thinks notes that "due to social media it will be increasingly difficult to write and enforce retail contracts with legal meanings very different from their 'common sense' meanings." That seems like important straightforward progress to me, but I suppose there is a Straussian reading. And what about smart contracts

Is whistle-blowing the new short selling?

There is a lot going on in this story about an investigation of AmTrust Financial Services Inc. The Securities and Exchange Commission is involved, the Federal Bureau of Investigation was involved, and at one point a former BDO USA LLP auditor secretly recorded conversations about AmTrust's audits on "a tiny recording device disguised as an ordinary Starbucks gift card" that was given to him by the FBI. (No charges have resulted, though AmTrust has a new auditor and has restated past results.)

But the most interesting part may be the involvement of the professional whistle-blowing industry. The BDO auditor went to the government to be a whistle-blower, but not directly: First he "joined forces in 2013 with a larger group that includes Harry Markopolos, a forensic accountant who warned the SEC about the Madoff scheme before it became public in late 2008." And "the Markopolos group hopes to profit by collecting a reward under the SEC’s Whistleblower Program, if the agency ever successfully brings legal action in the matter." 

Markopolos seems to have no direct inside knowledge of wrongdoing at AmTrust; his role is as an adviser, explainer, and hype man for the actual whistleblower, the BDO guy:

In a presentation that the group says it gave to the FBI and federal prosecutors in 2014, Mr. Markopolos’s team called one set of alleged accounting moves “The Washing Machine” for its purported cleansing effect on the bottom line, and another “The Loss Cemetery,” for its alleged effectiveness in burying losses in offshore affiliated entities.

The recent accounting restatement, with KPMG’s approval, “does not change our fundamental view of AmTrust’s shoddy accounting,” says Mr. Markopolos, who says his group is not betting against AmTrust’s stock.

Notice that last clause. It occurred to Markopolos to say (or to the reporter to ask him) that he's not shorting AmTrust stock. (He's betting against the company, sure, but not the stock.) Because of course the traditional way to take down a company that you think is a fraud is to short the stock and then go public with a big catchy presentation about how the company is a fraud. Hopefully the regulators will be convinced by your presentation, though it's mostly sufficient for your purposes if other investors are convinced.

The Markopolos approach is different, though there are some obvious common elements. There's still the catchy presentation, though his was given directly to regulators, not to the public. There's still the expectation of a big payday, though this comes from the SEC's Office of the Whistleblower, which can award 10 to 30 percent of any SEC sanctions to whistle-blowers, rather than from the market. And there's still a professional involved: Just as activist short selling works best if it's done by a professional hedge fund manager with a good track record, so whistle-blowing often seems to go through professional whistle-blowers like Markopolos with a track record and contacts at the SEC.

Is this ... better ... than short selling? Bill Ackman's activist short campaigns MBIA Inc. and Herbalife Ltd. have rubbed a lot of people the wrong way, even though he was right about MBIA and essentially correct about Herbalife. The complaint seems to be: What gives this rich hedge fund manager the right to try to get regulators to punish a business just so he can make money? Trying to make money on a business's failure -- or even on a fraud -- is just sort of unseemly, as is lobbying government regulators to do something that will make you money. Or at least, it's unseemly when it's done by short sellers. Is it more seemly when it's done by professional whistle-blowers? In some ways Ackman's Herbalife presentation -- given to the public, though obviously aimed in part at regulators -- seems like fairer play than Markopolos's AmTrust presentation, which was given confidentially to regulators, and which relies totally on convincing them, not the market. But then I was never bothered by activist short selling in the first place so I am probably the wrong person to ask.

Inscrutable AI.

Here is a delightful MIT Technology Review article by Will Knight about "The Dark Secret at the Heart of AI," which is that "the system is so complicated that even the engineers who designed it may struggle to isolate the reason for any single action":

Banks, the military, employers, and others are now turning their attention to more complex machine-learning approaches that could make automated decision-making altogether inscrutable. Deep learning, the most common of these approaches, represents a fundamentally different way to program computers. “It is a problem that is already relevant, and it’s going to be much more relevant in the future,” says Tommi Jaakkola, a professor at MIT who works on applications of machine learning. “Whether it’s an investment decision, a medical decision, or maybe a military decision, you don’t want to just rely on a ‘black box’ method.”

The computer figures out what to do -- how to drive, what stocks to pick -- and then does it. If you asked it why it picked the stocks it picked, it can't say "because they have low price-earnings ratios" or "because I read a nice story about them on Bloomberg this week" or "because I like the cut of the CEO's jib." It can only say "my model looks at a ton of inputs and draws subtle signals from them, which I use to produce outputs." "Which inputs did you use," you ask, and the computer replies: "All of them."

Knight views this as a big problem, and I suppose it is, but there's a sense in which it is inevitable. If an artificial intelligence program could explain its outputs in easily comprehensible heuristics, it wouldn't be useful: You could just use the heuristics yourself. The computer is supposed to come up with insights that you couldn't, that are too deep and subtle to summarize. "The explanations provided will always be simplified," writes Knight, "meaning some vital information may be lost along the way."

I am sure this is all very important in military and medical and car-driving applications, and in the philosophy of consciousness. In investing, you know, if you have a box that produces money when you press a button, you should probably just keep pressing the button and let someone else worry about the deep meaning of what the box is up to.

Still it is troubling. There is a lot of magic in investing, a lot of drawing pretty pictures on charts and expecting them to tell you the future. The naive view of this magic is that it is magic: The pictures have a self-operative reality; the "head and shoulders" is a supernatural spirit, telling you what your stock will do next, if you can only understand his language. The plausible, mature view is that the pictures encapsulate some real truths about human behavior: You can recast technical analysis in terms of behavioral psychology, explain resistance levels as loss aversion, and generally explain all your market signals in terms of plausible stories about humans. And you really should. You can't have much confidence in a pretty pattern just for its own sake. There is a lot of data out there. Butter production in Bangladesh was once famously well-correlated with the S&P 500 Index, but there's no reason to think it's predictive. Chief executive officer golf handicaps are also correlated with stock performance, and there is an easy human story to tell there about CEOs who spend too much time golfing and not enough time running their companies.

But deep-learning investing strategies that can't explain themselves lose that connection to plausible human stories. They feel almost like a step backwards: We thought we could explain markets in terms of human behavior, but with enough artificial intelligence maybe we can't. Maybe it's magic after all.

PCAOB scandal!

The Public Company Accounting Oversight Board, the regulator responsible for supervising auditors, is among the most boring of regulators, distinguished mostly by the fact that people sometimes pronounce its name -- PCAOB -- as "peek-a-boo," which is cute. But now it has a wild scandal! Oh wait no actually it has a pretty dull scandal:

Five KPMG LLP partners, including the head of its audit practice, were fired after the Big Four accounting firm improperly obtained information about which audits its regulator planned to inspect, the company said.

The company’s regulator, the Public Company Accounting Oversight Board, began investigating a leak discovered in February of its plans to inspect KPMG’s work.

PCAOB reviews a sample of each accounting firm's audits, "and the findings are widely seen as a report card for how the firms are performing and whether audit quality is getting better or worse." KPMG seems to have found out in advance which audits would be reviewed: "If an accounting firm found out which audits would be subject to inspection before it completed them, it could devote more time to those projects to ensure they pass muster with regulators." So it just did too good a job. Meh. Though obviously now the PCAOB should go and review some KPMG audits that weren't on the leaked list. If KPMG never showed up for them, because it knew they wouldn't be graded, that would be bad.

People are worried that people aren't worried enough.

On Monday Deutsche Bank Markets Research put out a note arguing that this is "the most boring year ever for credit." (The title is "The most boring year ever !!!") You can of course put a quiet-too-quiet reading on that: "in Europe, for instance, only 2006 and 2007 had more boring starts than 2017," and look how 2007 turned out. 

People are worried about unicorns.

The head of public relations at Uber Technologies Inc. is leaving the company, and wouldn't you? It's nice of another transportation company to fling itself headlong into a public relations debacle this week; that should distract from Uber and give her some cover as she leaves. I imagine she's been waiting for that for a while. Elsewhere, a guy who worked at Snap Inc. for three weeks is suing, claiming that Snap was overstating user growth numbers, and does the fact that he worked there for three weeks make his claim to be a whistleblower more or less plausible? And here is a story about Kenya's tech hub in Nairobi, the "Silicon Savannah" or, as I like to call it, the Enchanted Savannah.

People are worried about bond market liquidity.

Here is a fascinating note from Zoltan Pozsar at Credit Suisse Fixed Income Research about "Excess Reserves and Global Dollar Funding." One thing that I suppose is obvious to the dollar-funding cognoscenti, but that I hadn't really understood before reading this note, is that changes in U.S. money-market-fund regulation -- which have pushed hundreds of billions of dollars out of "prime" money market funds (which lend to banks) and into "government" money market funds (which invest in Treasury securities and similar safe government-backed assets) -- have reduced unsecured dollar funding of foreign banks but not domestic ones. U.S. banks can still get dollar funding from government money-market funds; it's just a little more indirect. The government funds lend their inflows to the Federal Home Loan Bank system (which "absorbed about $250 billion of inflows"), which then lends them to the banks:

Yet another advantage of U.S. banks over foreign banks is the ability of the former to tap six-month advances to fund HQLA portfolios (see Chain 2C). If a U.S. bank can issue fewer six-month CDs due to prime funds’ diminished presence in the CD market, it did not necessarily lose that funding if the FHLBs are lending more via advances on the back of government funds’ increased purchases of agency discount notes and floaters. But if foreign banks can issue fewer six-month CDs, their next port of call is the “outer rim” of funding markets – either the FX swap market or the debt capital market (see here).

Also: "Three-month U.S. dollar Libor is slowly becoming an inactive funding point."

Anyway there is much of interest there (on monetary policy, covered interest parity, etc.), but I put it in this section because of an oddity noted by Alexandra Scaggs: Microsoft Corp. apparently did its own little Operation Twist to accommodate money-market-fund reform. Pozsar:

Money fund reform delivered just over $200 billion in net new demand for bills from government funds. Roughly one fifth of this demand was accommodated by the corporate treasury department of Microsoft Corp., which sold $40 billion of bills and replaced them with longer-dated notes.

Scaggs notes that "a big buyer (or buyers) that didn’t fit into normal categories started showing up at auctions last year, and bidding up two-year notes," and suspects it might be Microsoft. There was artificially inflated demand for Treasury bills -- as a bunch of money was flowing into government money funds that could only invest in bills, not 2-year notes -- and an opportunity for someone to get paid for meeting that demand by selling bills and buying notes. In the olden days, you'd think that role -- providing liquidity to investors by selling them what they want and buying a closely-related substitute that they don't want -- would be taken by banks. Now, it's Microsoft.

Things happen.

Fed officials haven't stopped holding private meetings even after a criminal probe into leaks. China Regulator Warns Banks Away From Speculative Activity. Activist investor John Paulson to leave AIG board. Tesla Investors Press for Board Members Without Ties to Musk. Congressmen Warn Venezuela Default Could Lead to Russian Control of U.S. Oil Infrastructure. The money machine: how a high-profile corruption investigation fell apart. Swift Releases New Controls, Suspicious-Payments Blocker. ‘Genius’ Banker Randy Work Loses $225 Million Divorce Battle. JPMorgan Delays Shifting Some Wealth Clients as U.S. Mulls Rules. Trump promises again to revamp Wall Street reform rules. U.S. Unlikely to Tag China Currency Manipulator, Schwarzman Says. Brexit risks pushing UK out of European space contracts. Try not to be too good at your job. Tesla employees can't park. New Law to Compel Zimbabwe Banks to Accept Cattle as Collateral. Condom-clogged pipe provides police clue in cracking Texas brothel. 

If you'd like to get Money Stuff in handy email form, right in your inbox, please subscribe at this link. Thanks! 

This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.

    To contact the author of this story:
    Matt Levine at

    To contact the editor responsible for this story:
    James Greiff at

    Before it's here, it's on the Bloomberg Terminal.